09-18-2014 04:35 PM
@crossrulz wrote:
Another funny, but really sad, thread. A FINAL YEAR Electrical Engineering student doesn't know how to measure current. Heaven help us...http://forums.ni.com/t5/LabVIEW/Measuring-current-using-USB-6009-with-a-Rheostat-as-shunt/m-p/299646...
That's nothing. I had a Graduate student who reported to me, earning his masters, EE, who couldn't use a DC power supply. He had the "red" hooked up because he needed it to be +12V. He thought he didn't need the "black".
09-18-2014 06:28 PM
09-19-2014 08:02 AM
@crossrulz wrote:
Another funny, but really sad, thread. A FINAL YEAR Electrical Engineering student doesn't know how to measure current. Heaven help us...http://forums.ni.com/t5/LabVIEW/Measuring-current-using-USB-6009-with-a-Rheostat-as-shunt/m-p/299646...
"Your data are messed up. They look like they're off by a factor of 100."
"I used a 100 ohm shunt instead of 1 ohm. Just 'fix' it in the software. That's what you guys do, isn't it?"
"Sure ... ok ... that works ... but these numbers look a little high ... you better fix your hardware so we can get accurate data."
"DENIER! DENIER! DENIER! DENIER! DENIER! DENIER!"
09-19-2014 08:35 AM
"I used a 100 ohm shunt instead of 1 ohm. Just 'fix' it in the software. That's what you guys do, isn't it?"
I've been asked to "fix" measurements before. I'm sure I can't be the only one.
I can recall a customer once asking a collegue to add an "adjustable offset factor" that would basically allow a "failing" part to "pass."
You know, if all you wanted was a green a light; you could have done that yourself.
09-20-2014 06:47 PM
@pallen wrote:
[...] You know, if all you wanted was a green a light; you could have done that yourself.
It would certainly be easier for us to write. The Software Quality engineer would probably catch it, though.
09-20-2014 10:27 PM - edited 09-20-2014 10:29 PM
@jcarmody wrote:
@pallen wrote:[...] You know, if all you wanted was a green a light; you could have done that yourself.It would certainly be easier for us to write. The Software Quality engineer would probably catch it, though.
Sometimes they actually do:
<Sea Story Again>
Bare in mind the "code under test" in this case was proven good the Original validation plan was found to be questionable and replaced.
Validation test plan for a design verification test of a "Gizmo" "Gizmos" have an operating requirement to output a DC pulse if no input pulse was observed within a specific time. The verification test was to determine if the input reciever's sensitivity was per intended design. ( Of course it must have met its designed limits! All electronic systems perform as designed- some may have design flaws. I can't tell you how many times I've had to tell an engineer "It works just exactly the way you designed it to work")
Easy, Start but inputing simulated signals into the device, decrease signal amplitude to the minimum sensitivity threshhold and observe that at mininimum sensitivity Pulse output was still inhibited (must have met the sensing threshhold right?)
While I was a signator on the validation test report I did my due dilligance and actually inspected the test set-up. UUT out of loop !) the tech had simply forgotten to hook it up but of course- the validation passed since- no output was seen.
Procedures are in place to prevent that type of thing from ever recurring at that company.
</Sea Story>
09-21-2014 06:20 AM
Jeff·Þ·Bohrer wrote:the tech had simply forgotten to hook it up but of course- the validation passed since- no output was seen.
Ahhh, the classic "There was no response therefore we passed." I was doing a code walk through on a test system that I inherited and was told to update per the new specifications. It was an AM Rejection test: send a signal that is AM istead of FM and make sure the UUT doesn't see the command. In the walk through, I discovered that no modulation was being done. But there was no repsonse, so the UUT was good! Yeah, that opened up a huge can of worms since that code tested many different projects.
As for your sensitivity test, I would have written it so that you lower the voltage until you do see a response from the UUT. That would give you the real sensitivity instead of just making sure it met the exact spec.
09-22-2014 08:15 AM
@crossrulz wrote:
Jeff·Þ·Bohrer wrote:the tech had simply forgotten to hook it up but of course- the validation passed since- no output was seen.
As for your sensitivity test, I would have written it so that you lower the voltage until you do see a response from the UUT. That would give you the real sensitivity instead of just making sure it met the exact spec.
I would not have written the test that way either! Worm can WAS opened. Much code was re-validated. Many procedures were audited. Some DVT test were re-run on products available to the market. No design flaws of product were discovered. But we sweated bullets until we could prove designs. "Gizmos" were FDA class III Implantable devices.
10-24-2014 05:55 AM
This reminds me of what I think was a BOFH skit originally, but ended up being one of those 'too stupid to own a computer' chain emails about someone ringing tech support to complain about their PC not working in the middle of a power cut.
10-24-2014 08:37 AM
@thoult wrote:
This reminds me of what I think was a BOFH skit originally, but ended up being one of those 'too stupid to own a computer' chain emails about someone ringing tech support to complain about their PC not working in the middle of a power cut.
I actually had a customer (early 90's) return a PC TWICE because he did not want to change the "Invert Logic" switch setting. (You know, the big rocker switch in the back that has a "Zero" and a "One" on it )